It is already known to use appropriate sensors in a vehicle for occupant classification. Such sensors include, for example, a seat mat, present in the vehicle seat, which furnishes data on the basis of which a weight determination and also a seat position determination for the occupant are possible. These data furthermore make it possible to ascertain whether what is present is an occupant or an object or, in particular, a child seat. A further possibility is to use an optical sensor suite to perform the occupant classification. An occupant classification on the basis of high-frequency signals is also possible, exploiting the fact that an occupant possesses a dielectric constant different from that of air. Occupant classification is necessary for an airbag for which on the one hand a determination can then be made, based on the occupant classification, as to whether an airbag can be triggered at all without injuring the occupants; and in the case of an adjustably triggerable airbag, it is possible to decide which triggering stage is optimal for the particular occupants.
The apparatus according to the present invention for occupant classification in a vehicle has the advantage that the use of a CO2 sensor for occupant classification results in independence from environmental influences such as temperature and humidity. The occupant""s clothing and the use of seat cushions thus also have no influence on the weight determination. The use of a CO2 sensor also represents a low-cost solution, since a CO2 sensor can be manufactured extremely economically on a mass-produced basisxe2x80x94for example, as a resistive gas sensor and thus as a semiconductor sensor.
It is particularly advantageous that from the signals of the CO2 sensor it is possible on the one hand to determine the CO2 quantity (on the basis of which the weight and body area can be determined), and on the other hand to determine from the time signal of the CO2 sensor the respiratory frequency, in order thereby, in combination with the CO2 quantity, to perform a plausibility test of the respective parameters. Also possible on the basis of the respiratory frequency, for example, is a determination as to whether a small child is present, since he or she usually exhibits a higher respiratory frequency than an adult. The occupant classification accuracy, i.e. the distinction between child and adult, can thereby be further improved.
According to the present invention, the CO2 sensor is advantageously arranged very close to the occupant and also in front of him or her, so that exhaled air can be directly dynamically sensed by the CO2 sensor. A very accurate measurement is then therefore possible, since the processor can then make an accurate estimate on the basis of those data.
It is moreover advantageous that the CO2 sensor can be used in combination with a further sensor, for example with the seat mat, as a plausibility sensor, so as thereby to confirm or cast doubt on the signals of those other sensors.
It is furthermore advantageous that the processor cyclically polls the individual CO2 sensors allocated to the respective seats, and thus continuously has data from all the sensors.
It is advantageous in this context that the processor then classifies the individual occupants based on all the signals from all CO2 sensors. As a result, it can perform a plausibility check of the respective signals by way of the individual signals of the individual CO2 sensors. This can be accomplished, for example, using known pattern recognition methods, or by comparison with known data sets that have been generated, for example, in preliminary tests.